# How to use the .cross operator for 100 x 100 docking
**URL:** https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361
**Category:** Ask for help
**Tags:** nextflow
**Created:** [September 12, 2025, 9:09pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361 "2025-09-12T21:09:15Z")
**Posts on this page:** 13
**Page:** 1
### Author:  [@Funmilola\_Bashorun](https://community.seqera.io/u/Funmilola_Bashorun)
#### Post date: [September 12, 2025, 9:09pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/1 "2025-09-12T21:09:15Z")
Hi,
I am conducting virtual screening of some compounds and I am having issues with the docking phase. I wonder why the .cross operator is not generating any output for my ligand vs receptor pairs. I will appreciate any assistance.
Thank you.
---
### Author:  [@Alexander\_Nater](https://community.seqera.io/u/Alexander_Nater)
#### Post date: [September 16, 2025, 11:52am UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/2 "2025-09-16T11:52:36Z")
You need to provide a code example, otherwise it’s impossible to figure out what’s going on.
---
### Author:  [@Funmilola\_Bashorun](https://community.seqera.io/u/Funmilola_Bashorun)
#### Post date: [September 16, 2025, 1:15pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/3 "2025-09-16T13:15:44Z")
workflow {
```auto
// Step 0: Load ligand files
ligand_files = Channel.fromPath("${params.ligands}/\*.sdf", checkIfExists: true)
// Step 1: Split multi-ligand SDFs
split_sdf_files = splitLigand(ligand_files).flatten()
// Step 2: Rename by ZINC ID
renamed_sdf = renameSDF(split_sdf_files)
// Step 3: Minimize renamed ligands
minimized_sdf = ligandMinimization(renamed_sdf)
// Step 4: Convert minimized ligands
prepared_ligands = convertLigandtoPDBQT(minimized_sdf)
prepared_ligands_flat = prepared_ligands.flatten()
// Step 5: Load receptor files
receptor_files = Channel.fromPath("${params.receptors}/\*.{pdb,sdf,mol2}", checkIfExists: true)
// Step 6: Convert receptor files to .pdb format
converted_pdbs = convertReceptortoPDB(receptor_files)
// Step 7: Convert .pdb receptor files to .pdbqt
prepared_receptors = convertReceptortoPDBQT(converted_pdbs)
prepared_receptors_flat = prepared_receptors.flatten()
// Call docking process
receptor_centers = Channel.fromPath(‘results/receptor_centers.csv’)
.splitCsv(header:true)
.map { row ->
rec_file = file("results/prepared_pdbqt/${row.receptor}.pdbqt")
\[ rec_file, row.cx.toBigDecimal(),
row.cy.toBigDecimal(),
row.cz.toBigDecimal() \]
}
.filter { it\[0\].exists() } // skip missing files
// Step 9: Cartesian product of receptors and ligands
docking_pairs = prepared_receptors_flat.cross(prepared_ligands_flat)
// Step 10: Run docking
vina_docking(docking_pairs)
```
---
### Author:  [@bentsherman](https://community.seqera.io/u/bentsherman)
#### Post date: [September 16, 2025, 2:14pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/4 "2025-09-16T14:14:45Z")
The easiest way to debug this is to use the `view` operator to view the upstream channels. If the `cross` isn’t producing any output it’s likely because one of the input channels are empty.
---
### Author:  [@Alexander\_Nater](https://community.seqera.io/u/Alexander_Nater)
#### Post date: [September 17, 2025, 7:54am UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/5 "2025-09-17T07:54:25Z")
Why are you escaping the square brackets in your script (`\[`)?
---
### Author:  [@Funmilola\_Bashorun](https://community.seqera.io/u/Funmilola_Bashorun)
#### Post date: [September 19, 2025, 9:28pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/6 "2025-09-19T21:28:12Z")
The input channels give the required output. .cross is not generating the required output, I wonder if I am using it correctly.
---
### Author:  [@Funmilola\_Bashorun](https://community.seqera.io/u/Funmilola_Bashorun)
#### Post date: [September 19, 2025, 9:29pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/7 "2025-09-19T21:29:09Z")
That is probably due to the length of the code. It is without that on my end.
---
### Author:  [@bentsherman](https://community.seqera.io/u/bentsherman)
#### Post date: [September 22, 2025, 2:30pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/8 "2025-09-22T14:30:03Z")
In that case, can you give an example that just creates some fake input channels and calls the `cross` operator? That way I can run it for myself and see what’s wrong
---
### Author:  [@Funmilola\_Bashorun](https://community.seqera.io/u/Funmilola_Bashorun)
#### Post date: [September 23, 2025, 5:25pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/9 "2025-09-23T17:25:24Z")
#!/usr/bin/env nextflow
nextflow.enable.dsl=2
workflow {
```auto
ligands = Channel.of("lig1.pdbqt", "lig2.pdbqt", "lig3.pdbqt")
receptors = Channel.of(
["recA.pdbqt", 10.0, 20.0, 30.0],
["recB.pdbqt", 40.0, 50.0, 60.0]
)
pairs = ligands.cross(receptors)
```
}
---
### Author:  [@bentsherman](https://community.seqera.io/u/bentsherman)
#### Post date: [September 23, 2025, 5:55pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/10 "2025-09-23T17:55:25Z")
Okay I see. The caveat with the [cross](https://nextflow.io/docs/latest/reference/operator.html#cross) operator is that it only crosses inputs that have a matching key. Despite it’s name, it does not do a full cross product.
You can specify a closure to define the matching key in terms of an input value. So you could make this closure return a constant value to force a full cross product:
```auto
ligands.cross(receptors) { v -> 0 }
```
Honestly this should probably be the default behavior, but this was before my time 🤷♂️
---
### Author:  [@Funmilola\_Bashorun](https://community.seqera.io/u/Funmilola_Bashorun)
#### Post date: [September 23, 2025, 7:10pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/11 "2025-09-23T19:10:57Z")
PAIR: [lig1.pdbqt, recA.pdbqt]
PAIR: [lig1.pdbqt, recB.pdbqt] - it does not ouput all pairs. Is there another way to achieve a cartesian product for 50 x 50 docking?
#!/usr/bin/env nextflow
nextflow.enable.dsl=2
workflow {
```auto
ligands = Channel.of("lig1.pdbqt", "lig2.pdbqt", "lig3.pdbqt")
receptors = Channel.of(
["recA.pdbqt", 10.0, 20.0, 30.0],
["recB.pdbqt", 40.0, 50.0, 60.0]
)
docking_pairs = receptors.flatMap { receptor_list ->
ligands.map { ligand ->
[
receptor_list[0], // receptor file
receptor_list[1], // cx
receptor_list[2], // cy
receptor_list[3], // cz
ligand // ligand file
]
}
}
docking_pairs.view { p -> "DOCK_PAIR: ${p}" }
```
}
\>\> DOCK\_PAIR: DataflowBroadcast around DataflowStream[?]
DOCK\_PAIR: DataflowBroadcast around DataflowStream[?]
---
### Author:  [@bentsherman](https://community.seqera.io/u/bentsherman)
#### Post date: [September 23, 2025, 8:42pm UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/12 "2025-09-23T20:42:38Z")
If you want the outputs to be flat, you can use the `combine` operator:
```auto
ligands.combine(receptors)
```
---
### Author:  [@system](https://community.seqera.io/u/system)
#### Post date: [November 22, 2025, 12:16am UTC](https://community.seqera.io/t/how-to-use-the-cross-operator-for-100-x-100-docking/2361/13 "2025-11-22T00:16:26Z")
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.